Search

CN-122022571-A - Talent assessment system and method based on multi-source data fusion and intelligent analysis

CN122022571ACN 122022571 ACN122022571 ACN 122022571ACN-122022571-A

Abstract

The invention relates to a talent assessment system and a talent assessment method based on multi-source data fusion and intelligent analysis, belonging to the crossing field of information technology and human resource management. The technical problem to be solved is that multisource heterogeneous data are difficult to fuse, an evaluation model is static and stiff, decisions lack data support, and closed loop self-optimization cannot be achieved in existing talent evaluation. The technical scheme is characterized in that multiple source heterogeneous data such as resume, interview score, background survey and the like in a recruitment whole process are subjected to cleaning and standardization processing through data acquisition and structuring steps and are converted into unified structured data, double quantitative calculation of a process stage quantization model and a capacity dimension quantization model is executed in parallel based on the structured data, process stage linear score and multi-dimensional capacity radar map score of a candidate are respectively generated, the dual model score is integrated, a comprehensive evaluation result and a visual report are generated according to a preset decision rule, performance data of recorded staff are acquired, correlation between historical evaluation and performance is analyzed through a machine learning algorithm, and model weight parameters are dynamically adjusted, so that feedback and self-optimization of the system are realized. The method is mainly used for enterprise recruitment, talent selection and performance prediction, and can improve the objectivity, the accuracy and the decision efficiency of evaluation.

Inventors

  • Xie Hongkang
  • LIU YING

Assignees

  • 北京银证大通国际咨询有限公司

Dates

Publication Date
20260512
Application Date
20260126

Claims (9)

  1. 1. A recruitment process talent assessment method based on multi-source data fusion is characterized by comprising the steps of collecting multi-source heterogeneous data in a recruitment process, cleaning and standardizing the multi-source heterogeneous data, converting the multi-source heterogeneous data into unified structured data and storing the unified structured data, performing process stage quantization model calculation and capacity dimension quantization model calculation in parallel based on the structured data, wherein the process stage quantization model calculation is used for generating process stage linear scores of candidates, the capacity dimension quantization model calculation is used for generating multi-dimensional capacity radar chart scores of the candidates, the data fusion and comprehensive assessment step is used for fusing the process stage linear scores with the multi-dimensional capacity radar chart scores and generating comprehensive assessment results according to preset decision rules, and the model feedback and self-optimizing step is used for collecting performance data of recorded staff, analyzing the historical evaluation data and performing correlation of performance weight data through a machine learning algorithm, and adjusting dynamic parameters of the dynamic quantization model.
  2. 2. The method of claim 1, wherein the process phase quantization model calculation includes defining a plurality of successive recruitment phases, each phase having a full value and a weight, according to a formula , Calculate a total flow score S process, and generate a visual flow score map.
  3. 3. The method of claim 1, wherein the capability dimension quantization model calculation comprises predefining a plurality of capability dimensions, the capability dimensions comprising a set of out-of-training dimensions and a set of in-training dimensions, according to a formula A weighted average score S ability(j), for each capability dimension is calculated and the multi-dimensional capability radar map is generated.
  4. 4. The method of claim 3, wherein the set of out-training dimensions includes post expertise capability, analysis decision capability, plan execution capability, team collaboration capability, and self-management capability, and wherein the set of in-training dimensions includes insight capability, internal driving force, persistence capability, tank integrity, and integrity.
  5. 5. The method of claim 1, wherein in the step of data fusion and comprehensive evaluation, the predetermined decision rule comprises automatically marking a candidate as recommended recording when a total flow score S process is greater than a threshold T1 and a score S ability(j) of at least one predefined critical capability dimension is greater than a threshold T2.
  6. 6. The full-flow talent assessment system for implementing the method according to any one of claims 1 to 5, characterized in that the system adopts a layered architecture, and comprises a data acquisition and interaction layer for acquiring multi-role input original data through a Web front end and a mobile APP interface, a data processing and service logic layer which is a core layer and comprises a data cleaning module, a standardization module and a model calculation engine for executing the double quantification model calculation, a data storage layer for storing the processed structured data and unstructured original data, and an intelligent service layer for providing a reusable algorithm service interface comprising a weight calculation algorithm and a model self-optimization algorithm.
  7. 7. The system of claim 6, wherein the data processing and business logic layer further comprises a data parsing module for parsing a resume file and extracting structured information, a natural language processing module for processing text comments and background survey texts and performing keyword extraction and emotion analysis, and a data fusion module for performing the data fusion and comprehensive evaluation steps and generating a visual report.
  8. 8. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the method of any one of claims 1 to 5 when the program is executed by the processor.
  9. 9. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the method according to any one of claims 1 to 5.

Description

Talent assessment system and method based on multi-source data fusion and intelligent analysis Technical Field The invention relates to the crossing field of information technology and human resource management, in particular to a talent assessment system and method based on multi-source data fusion and intelligent analysis, which are suitable for the scenes of enterprise recruitment, talent selection, performance prediction and the like. Background Current talent assessment relies mainly on two types of techniques: The recruitment management system (ATS) based on the relational database has the core functions of flow management (such as resume screening and interview arrangement), but an evaluation module only supports simple form scoring, data isolation and lack of deep analysis capability, and the on-line psychological assessment tool can generate a standardized report, but has the following defects that the model is stiff and is difficult to fuse with recruitment flow data in the prior art: 1. the data isomerism is prominent, namely, the multi-source data formats such as resume, interview score, background investigation and the like are different, and the systems cannot be communicated with each other to form an information island; 2. the assessment model is static, namely, the grading dimension of the ATS and an online assessment algorithm are fixed, and the weight cannot be dynamically adjusted according to the post requirements; 3. The decision support is weak, the cross-stage association analysis is lacking, and the association of the index and the future performance cannot be evaluated through historical data mining; 4. And the closed loop learning is missing, namely, the system does not introduce feedback of performance data after employee attendance, and an evaluation model cannot be optimized. Disclosure of Invention 1. Technical problem to be solved The invention breaks through the bottleneck of the prior art and solves the problems of realizing unified fusion and structuring processing of multi-source heterogeneous data in the recruitment whole process, providing a dynamically configurable quantitative evaluation model, supporting self-adaptive adjustment according to post requirements and providing predictive support for decision making through data mining and visual analysis. 2. Technical proposal The core architecture of the system comprises a data acquisition layer, a business logic layer, a storage layer and an intelligent service layer, and the key innovation is as follows: 1. The multi-source data fusion technology comprises the steps of analyzing unstructured data (such as interview comments and background investigation texts) through Natural Language Processing (NLP), extracting keywords and converting the keywords into structural indexes, constructing a unified data model taking a candidate ID as a main key, and associating heterogeneous data such as resume, grading, operation and the like. 2. Double quantization dynamic model: Defining N continuous recruitment stages (such as initial trial, final trial and back adjustment), calculating scores of each stage, then weighting and summing to generate a linear flow scoring graph, wherein the formula is as follows: ; the capacity dimension quantization model is used for presetting a dimension set of 'external training' (such as professional skill) and 'internal training' (such as internal driving force), aggregating cross-stage scoring of each dimension to generate a radar map, wherein each dimension scoring formula is as follows: 。 3. And the closed loop self-optimizing mechanism is that the system periodically collects performance data (such as KPI) of the staff for recording, analyzes the relevance between the historical scores and the performance through a machine learning algorithm (such as linear regression) and automatically adjusts the weight parameters of the model. 3. Advantageous effects 1. The technical effects are as follows: Breaking the information island through a multi-source data fusion technology to form a panoramic view of the candidate; The double-model dynamic configuration enables the evaluation result to be more fit with the actual demands of the posts, and the accuracy is improved; The subjective bias is reduced based on data-driven predictive analysis, and the decision scientificity is remarkably enhanced. 2. Commercial value: recruitment efficiency is improved, namely a visual report is automatically generated, and a decision period is shortened; Drop-in risk reduction, namely deep mining of potential characteristics of the candidate through the capacity dimension radar chart; and the talent data assets of enterprises are accumulated, and support is provided for echelon construction. Drawings FIG. 1 is a schematic diagram of the overall architecture of the system showing the interaction relationship of data acquisition, processing, storage and service layers; FIG. 2 is a flow chart of a dual quantization model ca